|
Study Limitations
<br />All public opinion research is subject to unmeasured error. While the methodologies employed for this survey were
<br />designed to minimize this error as much as possible, these other sources of potential error should be
<br />acknowledged, and can include non-response error, coverage error, recall bias and social desirability bias.
<br />Non-response error arises when those who were selected to participate in the survey did not do so, and may have
<br />different opinions or experiences that survey responders. For general resident surveys, where the results are meant
<br />to be generalized to the entire adult population living in households, the mailing lists based on the Delivery
<br />Sequence File from the United States Post Office may exclude certain types of housing units, such as those in
<br />multi-family buildings where mail is addressed to a named resident at the address rather than to a specific unit or
<br />where residents only receive their mail at a post office box and the geographic location of a residence cannot be
<br />determined, there may be a coverage error , although for most locations, this is minimal. Respondents may not
<br />perfectly remember their experiences in the past year (such as participation in social or civic events, for example),
<br />and for some survey items, they may answer in ways they think cast their responses in a more favorable light
<br />(recall bias and social desirability bias).
<br />Survey Validity
<br />See the Polco Knowledge Base article on survey validity at https://blog.polco.us/polco-knowledgebase/margin-of-
<br />error-and-confidence-level-in-survey-results
<br />Contact
<br />Redwood City funded this research. Please contact Jennifer Yamaguma of Redwood City at
<br />JYamaguma@redwoodcity.org if you have any questions about the survey.
<br />1. See AAPOR's Standard Definitions for more information at https://aapor.org/standards-and-ethics/standard-definitions/
<br />2. Pasek, J. (2014). ANES Weighting Algorithm. Retrieved from https://surveyinsights.org/wp-content/uploads/2014/07/Full-anesrake-paper.pdf
<br />3. Targets come from the 2020 Census and 2023 American Community Survey
<br />Unweighted Weighted Target ⁴
<br />Age 18-34
<br />35-54
<br />55+
<br />Area 1
<br />2
<br />3
<br />4
<br />5
<br />6
<br />7
<br />Hispanic
<br />origin
<br />No
<br />Yes
<br />Housing
<br />tenure
<br />Own
<br />Rent
<br />Housing type Attached
<br />Detached
<br />Race &
<br />Hispanic ori..
<br />Not white alone
<br />White alone, not Hispanic or Latino
<br />Sex Man
<br />Woman
<br />Sex/age Man 18-34
<br />Man 35-54
<br />Man 55+
<br />Woman 18-34
<br />Woman 35-54
<br />Woman 55+
<br />weighting were age, sex, race, Hispanic origin, housing type, housing tenure, and area. No adjustments were made
<br />for design effects. Weights were calculated using an iterative, multiplicative raking model known as the ANES
<br />Weighting Algorithm.² The results of the weighting scheme for the probability sample are presented in the following
<br />table.
<br />Polco aligns demographic labels with those used by the U.S. Census for reporting purposes, when possible. Some
<br />categories (age, race/Hispanic origin, housing type, and length of residency) are combined into smaller subgroups.
<br />Methods
<br />Selecting Survey Recipients
<br />All households within Redwood City were eligible to participate in the survey. A list of all households within the zip
<br />codes serving Redwood City was purchased from Polco's mailing vendor, based on updated listings from the United
<br />States Postal Service.
<br />Since some of the zip codes that serve Redwood City households may also serve addresses that lie outside of the
<br />community, the exact geographic location of each housing unit was compared to community boundaries using the
<br />most current municipal boundary file. Addresses located outside of Redwood City boundaries were removed from the
<br />list of potential households to survey. Each address identified as being within City boundaries was further identified
<br />as being within one of the seven areas. From that list, addresses were randomly selected as survey recipients, with
<br />multi-family housing units sampled at a rate of 5:3 compared to single family housing units.
<br />Conducting the Random Sample Survey
<br />The 3000 randomly selected households received mailings beginning on September 4, 2025 and data collection for
<br />the survey remained open for eleven weeks. The first mailing was a postcard inviting the household to participate in
<br />the survey online. The next mailing contained a cover letter with instructions, the survey questionnaire, and a
<br />postage-paid return envelope. The final mailing was a reminder postcard giving the household one last opportunity to
<br />participate in the survey. All mailings included a web link to give residents the opportunity to respond to the survey
<br />online, as well as QR codes to further encourage participation.
<br />About 0.2% of the 3000 mailed invitations were returned because the household address was vacant or the postal
<br />service was unable to deliver the survey as addressed. Of the remaining 2,994 households that received the
<br />invitations to participate, 280 completed the survey, providing an overall response rate of 9%. The response rate was
<br />calculated using AAPOR’s response rate #2 for mailed surveys of unnamed persons.¹
<br />It is customary to describe the precision of estimates by a “level of confidence” and accompanying “confidence
<br />interval” (or margin of error). A traditional level of confidence, and the one used here, is 95%. The 95% confidence
<br />interval quantifies the sampling error or imprecision of the survey results based on the total number of responses
<br />received. This is because some residents’ opinions are relied on to estimate all residents’ opinions. The margin of
<br />error for the Redwood City survey is no greater than plus or minus 5.9 percentage points around any given percent
<br />reported for all respondents (280 completed surveys).
<br />Conducting the Open Participation Survey
<br />In addition to the randomly selected “probability sample” of households, a link to an online open-participation survey
<br />was publicized by Redwood City. The open-participation survey was identical to the random sample survey, with two
<br />small updates; it asked a question to confirm the respondent was a resident of Redwood City and also a question
<br />about where they heard about the survey.
<br />The open-participation survey was open to all city residents and became available on October 2, 2025. The survey
<br />remained open for seven weeks and 290 responses were received. The data presented in the following report
<br />excludes the open participation survey data, but the online report includes a tab which provides the complete open
<br />participation results.
<br />Analyzing the Data
<br />Responses from mailed surveys were entered into an electronic dataset using a “key and verify” method, where all
<br />responses are entered twice and compared to each other. Any discrepancies were resolved in comparison to the
<br />original survey form. Range checks as well as other forms of quality control were also performed. Responses from
<br />surveys completed on Polco were downloaded and merged with the mailed survey responses.
<br />The survey datasets were analyzed using all or some of a combination of the Statistical Package for the Social
<br />Sciences (SPSS), R, Python, and Tableau.
<br />The demographics of the survey respondents were compared to those found in the 2020 Census and 2023 American
<br />Community Survey estimates for adults in Redwood City. The primary objective of weighting survey data is to make
<br />the survey respondents reflective of the larger population of the community. The characteristics used for 3
<br />11.A. - Page 12 of 43
<br />146
|